Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Smart review router. Detects context (morning, Sunday, end of month) and launches the appropriate review workflow. Use anytime for the right review at the right time.
.claude/skills/ballred-review/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 104% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 23% | 0% |
Smart router that detects context and launches the appropriate review workflow.
/review # Auto-detect the right review based on time/context
/review daily # Force daily review
/review weekly # Force weekly review
/review monthly # Force monthly reviewOr simply: "Help me review" — and the right workflow starts.
When invoked without arguments, detect context using these rules:
bashHOUR=$(date +%H)
/daily morning workflow/daily evening workflow/daily midday workflowbashDAY_OF_WEEK=$(date +%u) # 1=Monday, 7=Sunday
/weeklybashDAY_OF_MONTH=$(date +%d) DAYS_IN_MONTH=$(date -v+1m -v1d -v-1d +%d 2>/dev/null || date -d "$(date +%Y-%m-01) +1 month -1 day" +%d)
/monthlyBefore routing, check for overdue reviews:
bash# Read weekly review file for last date WEEKLY_REVIEW="Goals/3. Weekly Review.md" # If last weekly review > 7 days ago, suggest weekly regardless of day
After detecting context:
This skill does NOT duplicate the logic of /daily, /weekly, or /monthly. It:
If the user specifies a type (/review weekly), skip auto-detection entirely and go directly to that review type.
markdown### Review Router **Time:** 7:15 AM (Morning) **Day:** Sunday **Month day:** 15th **Detected:** Weekly review (Sunday override) **Last weekly review:** 3 days ago (not overdue) Launching weekly review...
/review monthly runs monthly review even on a Tuesday morningWorks with:
/daily — Morning, midday, and evening routines/weekly — Full weekly review process/monthly — Monthly review and planning| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 4,964 | 7,334 | +48% | 1 | 1 | 0% | 724 | 1,139 | +57% | 0 | 0 | — |
case-02 | fail→fail | 8,679 | 11,981 | +38% | 1 | 1 | 0% | 1,295 | 2,147 | +66% | 0 | 0 | — |
case-03 | fail→fail | 5,722 | 7,927 | +39% | 1 | 1 | 0% | 876 | 1,461 | +67% | 0 | 0 | — |
case-04 | pass→pass | 6,219 | 2,472 | -60% | 1 | 1 | 0% | 1,138 | 1,333 | +17% | 0 | 0 | — |
case-05 | fail→pass | 7,408 | 3,959 | -47% | 1 | 1 | 0% | 1,166 | 1,621 | +39% | 0 | 0 | — |
case-06 | fail→pass | 9,594 | 2,974 | -69% | 1 | 1 | 0% | 1,513 | 1,452 | -4% | 0 | 0 | — |
case-07 | pass→pass | 4,722 | 3,573 | -24% | 1 | 1 | 0% | 773 | 1,528 | +98% | 0 | 0 | — |
case-20 | pass→pass | 10,485 | 8,736 | -17% | 1 | 1 | 0% | 1,762 | 2,292 | +30% | 0 | 0 | — |
case-08 | fail→fail | 8,338 | 3,283 | -61% | 1 | 1 | 0% | 1,252 | 1,493 | +19% | 0 | 0 | — |
case-09 | pass→pass | 5,816 | 3,558 | -39% | 1 | 1 | 0% | 827 | 1,511 | +83% | 0 | 0 | — |
case-10 | pass→pass | 8,459 | 4,227 | -50% | 1 | 1 | 0% | 1,315 | 1,745 | +33% | 0 | 0 | — |
case-11 | pass→pass | 11,506 | 4,508 | -61% | 1 | 1 | 0% | 2,106 | 1,831 | -13% | 0 | 0 | — |
case-12 | pass→pass | 4,538 | 4,861 | +7% | 1 | 1 | 0% | 793 | 1,814 | +129% | 0 | 0 | — |
case-13 | pass→pass | 8,565 | 3,698 | -57% | 1 | 1 | 0% | 1,306 | 1,570 | +20% | 0 | 0 | — |
case-14 | pass→pass | 3,684 | 1,917 | -48% | 1 | 1 | 0% | 538 | 1,206 | +124% | 0 | 0 | — |
case-15 | fail→fail | 8,453 | 3,256 | -61% | 1 | 1 | 0% | 1,395 | 1,375 | -1% | 0 | 0 | — |
case-16 | pass→pass | 7,203 | 2,823 | -61% | 1 | 1 | 0% | 1,022 | 1,389 | +36% | 0 | 0 | — |
case-17 | fail→pass | 6,916 | 3,009 | -56% | 1 | 1 | 0% | 1,128 | 1,462 | +30% | 0 | 0 | — |
case-18 | fail→pass | 4,932 | 4,183 | -15% | 1 | 1 | 0% | 845 | 1,722 | +104% | 0 | 0 | — |
case-19 | fail→pass | 8,944 | 5,138 | -43% | 1 | 1 | 0% | 1,566 | 1,920 | +23% | 0 | 0 | — |
case-21 | pass→pass | 8,582 | 5,913 | -31% | 1 | 1 | 0% | 1,696 | 2,056 | +21% | 0 | 0 | — |
case-22 | pass→pass | 9,370 | 7,887 | -16% | 1 | 1 | 0% | 1,758 | 2,322 | +32% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +23 percentage points is the difference between those two pass rates over the 19 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.